R-multiple is the return on a trade expressed as a multiple of the amount risked. A 2R win returned twice your initial risk; a −1R loss lost exactly what you risked. R normalizes performance across position sizes, leverage, and asset classes, which makes 100 trades comparable apples-to-apples even when account size and leverage change across them. The metric was popularized by Van Tharp in Trade Your Way to Financial Freedom (1999) and is the foundation of most modern position-sizing systems.
- R-multiple = (exit price − entry price) ÷ (entry price − stop price). Sign flipped for short trades. The number tells you the return as a multiple of initial risk.
- A 2R trade returned twice the amount you risked. A −1R trade lost exactly the amount you risked. Bigger R, bigger reward; the unit is consistent across trades.
- Expectancy = (win rate × average winning R) − (loss rate × average losing R). Profitable systems typically need expectancy ≥ 0.3R after costs.
- R is broker-agnostic, leverage-agnostic, and position-size-agnostic. That's why it works the same in forex, equities, futures, and crypto perpetuals.
- Win rate alone is meaningless. A 90% win rate with 0.2R winners and 5R losers has negative expectancy.
- For crypto perp traders specifically: R requires a hard stop. Without one, "initial risk" is undefined and the math falls apart.
What is R-multiple?
- R-multiple
- The return on a trade expressed as a multiple of the amount risked on that trade. R is a normalized unit — a 2R trade always means "twice the risk taken" regardless of whether you risked $100 or $10,000. Coined by Van Tharp in 1999 and now standard in serious trading-system design.
R-multiple solves a specific problem: dollar P&L is incomparable across trades because position sizes change. A $500 win on a $100-risk trade is excellent (5R). A $500 win on a $5,000-risk trade is mediocre (0.1R). Reading them as raw dollar amounts hides the difference. Reading them as R-multiples makes the difference immediate.
The same logic applies across asset classes. A 2R trade in BTC perps and a 2R trade in ES futures and a 2R trade in EUR/USD are directly comparable as performance — even though the dollar amounts, leverage, and volatility profiles are completely different. That portability is the entire point of the metric.
How do I calculate R-multiple?
Three steps. The math is intentionally simple because the math being simple is what makes the metric usable in real time.
- Define your initial risk (1R) before the trade. 1R = the dollar distance from your entry price to your stop-loss price, multiplied by your position size. If you enter long BTC at $60,000 with a $59,000 stop and 0.1 BTC of size, your 1R = $1,000 × 0.1 = $100.
- Record your actual P&L when the trade closes. Whatever the trade actually made or lost in dollars — your broker statement is the authoritative source, including fees and funding.
- Divide actual P&L by 1R. If 1R was $100 and the trade closed +$280, the R-multiple is 2.8R. If it closed −$100, the R-multiple is −1R. If you got stopped out for slightly worse than your planned stop (slippage), it might be −1.1R. If you exited early for a small win, it might be 0.4R.
Two practical notes:
- Define 1R from your planned stop, not the actual exit price. The whole purpose of R is to measure outcome relative to the risk you committed to at entry. If you moved your stop mid-trade, that's a separate analysis (and usually a journaling note worth writing down).
- Include fees and funding in the actual P&L. For crypto perps especially, funding rates can flip the sign of a small-R trade. Excluding them produces a flattering R-multiple that doesn't match what hit your account.
What is a good R-multiple?
There is no single "good" R-multiple — the right answer depends on your win rate. The two metrics are inseparable.
A scalping strategy might have a 70% win rate with average wins of 0.6R and average losses of −1R. A trend-following strategy might have a 35% win rate with average wins of 3R and average losses of −1R. Both can be profitable. Both can be unprofitable. The R-multiple in isolation tells you nothing without the win rate attached.
The honest target is expectancy, not R-multiple by itself:
Expectancy = (win rate × average winning R) − (loss rate × average losing R)
Worked examples:
| System | Win rate | Avg win R | Avg loss R | Expectancy |
|---|---|---|---|---|
| High win-rate scalp | 70% | 0.6R | −1R | (0.7 × 0.6) − (0.3 × 1) = 0.12R |
| Balanced swing | 50% | 1.5R | −1R | (0.5 × 1.5) − (0.5 × 1) = 0.25R |
| Trend-following | 35% | 3R | −1R | (0.35 × 3) − (0.65 × 1) = 0.40R |
| "Looks profitable" but isn't | 90% | 0.2R | −5R | (0.9 × 0.2) − (0.1 × 5) = −0.32R |
The last row is the trap. A 90% win rate feels like the best strategy at the table. The math says it bleeds.
What's the difference between R-multiple and win rate?
Win rate tells you how often you win. R-multiple tells you how much you win when you win, and how much you lose when you lose. Neither is meaningful in isolation; together they describe a strategy's expectancy.
The single most common retail mistake is optimizing for win rate. A trader sees their win rate drop from 65% to 55% after changing entries and concludes the new entries are worse. But if the average winning R rose from 0.8R to 1.6R at the same time, the new system is dramatically more profitable. The win-rate-only view missed this entirely.
The corollary: be skeptical of any trading strategy that markets itself on win rate alone. "85% win rate" tells you nothing about whether the strategy makes money. It often correlates inversely with profitability — selling cheap optionality for premium routinely produces 80%+ win rates with catastrophic average losses.
What is expectancy, and how does R fit in?
Expectancy is the average R-multiple a strategy produces per trade. It's the single most important number in trading-system evaluation. R-multiple is the unit; expectancy is the average.
Mechanically: take every closed trade, compute its R-multiple, average them. That's expectancy. A strategy with expectancy of +0.4R earns an average of 0.4R per trade across all trades (winners and losers combined). At 100 trades per year, that's 40R of total return. With each R sized at 1% of account, that's a 40% annual return before compounding.
Why expectancy beats win rate, profit factor, and Sharpe ratio as a primary metric:
- It's position-size-independent. A trader who doubles their per-trade risk doubles their dollar P&L without changing expectancy. That makes expectancy a clean measure of strategy quality, separate from sizing decisions.
- It's account-size-independent. Two traders with $5,000 and $500,000 accounts running the same strategy have the same expectancy. Their dollar outcomes differ; their strategy quality doesn't.
- It composes additively. Multiple strategies' expectancies can be combined (weighted by trade frequency) to compute portfolio-level expectancy. Win rate and profit factor don't compose cleanly.
A useful rule of thumb: any trading system that runs negative expectancy across 100+ trades is broken. Variance can mask negative expectancy for 20–30 trades; it cannot survive 100. If a strategy you've traded 100+ times shows expectancy below 0, the conclusion is not "I got unlucky" — it's "this isn't an edge."
How does R-multiple work in crypto perpetual futures?
The math is identical. The execution is harder, because perpetuals create two specific R-multiple problems that don't exist in spot or traditional futures.
First: hard stops are required, and many crypto traders don't use them. R-multiple depends on having a known 1R at entry. If a trader enters with no defined stop ("I'll get out when it looks wrong"), 1R is undefined and the math breaks. The typical workaround — measuring R after the fact based on where the trader actually exited — produces a flattering number that doesn't reflect risk taken. Real R requires a real, pre-committed stop.
Second: funding can flip the sign of small-R trades. On 8-hour funding intervals, a perp trader holding a position through three funding payments at 0.05% each can pay 0.15% of position size in funding alone. On a +0.4R trade, that funding cost might be the difference between +0.4R and +0.1R. The R-multiple recorded should reflect actual P&L net of funding — most journaling tools that auto-sync from broker APIs handle this automatically; manual entry usually misses it.
For traders new to perp trading, the two corrections are simple:
- Set a hard stop in the broker UI before clicking entry. Not "I'll watch it" — a literal stop order. Without that, R is theoretical.
- Use closed-position P&L (net of fees and funding) for R-multiple math, not high-water-mark unrealized P&L. The metric measures what you actually kept.
R-multiple in different asset classes
R-multiple is the same calculation everywhere. The only thing that changes is what counts as "1R" — i.e., how the initial-risk distance is measured.
| Asset class | 1R definition | Common pitfall |
|---|---|---|
| Equities (long) | Position size × (entry − stop) | Forgetting to subtract commission + slippage from R |
| Options (long premium) | Premium paid (max loss) | Treating the option price like a stop — 1R is usually the full premium |
| Forex | Pips × pip-value × position size | Misreading lot sizes and pip values across currency pairs |
| Futures (e.g., ES, NQ) | Tick distance × tick value × contracts | Treating margin as risk — margin ≠ 1R |
| Crypto spot | Position size × (entry − stop) | Forgetting exchange fees in net P&L |
| Crypto perpetuals | Position size × (entry − stop), with funding accrual netted in actual P&L | Trading without a hard stop; ignoring funding |
The unit translates cleanly because R-multiple is just a ratio. Every trade ever taken in any market, at any size, in any currency, can be expressed in R and compared directly to every other trade.
R-multiple, expectancy, and grade caps — auto-calculated.
Trade Journal AI computes R per trade and rolling expectancy automatically from your broker feed. Connect Bitunix, Hyperliquid, or Binance — no math required.
Frequently asked questions
Who invented R-multiple?
Van Tharp popularized R-multiple as a system-design tool in his book Trade Your Way to Financial Freedom (1999). The underlying concept — measuring return relative to risk taken — predates the book, but Van Tharp's framing as a normalized per-trade unit became the standard usage in trading-system literature.
Can R-multiple be greater than 10?
Yes, and routinely on trend-following strategies. A 15R or 20R outlier trade is uncommon but expected — these strategies often rely on a small number of large-R winners to offset many small-R losers. If a strategy never produces >5R winners but does produce −1R losers, expectancy struggles.
How is R-multiple different from risk/reward ratio?
Risk/reward ratio is the planned ratio at entry — e.g., "I'm risking 1R to make 3R, so this is a 3:1 trade." R-multiple is the realized outcome — what the trade actually delivered. Risk/reward is a forecast; R-multiple is a result.
Should I use R-multiple or percent return per trade?
R-multiple for strategy evaluation, percent return for account-level performance. The two answer different questions: R tells you whether your strategy has edge; percent tells you what you actually earned on your capital. Both matter.
What R-multiple should I target on my winning trades?
The wrong question. You don't target R — you target expectancy. A 1.5R average winner is fine if your win rate supports it. The right framing is: "What expectancy does my strategy produce, after costs, over 100+ trades?" If that number is positive and stable, the average winning R is whatever it is.
How do I calculate R-multiple for a trade I scaled out of in pieces?
Weighted average the exit price across your partials, treat that as your effective exit, and compute R against your original 1R. Most journal apps handle this automatically when they auto-sync from the broker; the broker reports each fill separately, the journal consolidates them into one weighted-average position.
Why does my R-multiple come out different in different journal apps?
Usually because one app is computing R from your stop-loss order and another is computing it from a "risk per trade" field you entered manually. Use the stop-loss method — it reflects actual risk at entry. If you don't use a hard stop, the manual-risk-per-trade method is the only option, but it requires discipline to keep accurate.
Does R-multiple work for swing trades held for weeks?
Yes — R is time-agnostic. A 3R swing trade held for 3 weeks and a 3R scalp closed in 3 minutes are equivalent on the R scale. Time enters the analysis at a different level (R per day, or trade frequency), not inside the R calculation itself.